
import numpy as np
import pandas as pd
import seaborn as sns 
import matplotlib.pyplot as plt
%matplotlib inline
from sklearn import tree
from sklearn.tree import DecisionTreeClassifier
from sklearn.metrics import accuracy_score
from sklearn.metrics import classification_report
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score


df = pd.read_csv('Visa_Applications_Agri.csv')


df.head()


df.shape

(17753, 35)


df.describe()


df.columns

Index(['CASE_NUMBER', 'CASE_STATUS', 'DAYS', 'TYPE_OF_EMPLOYER_APPLICATION',
       'H2A_LABOR_CONTRACTOR', 'NATURE_OF_TEMPORARY_NEED', 'EMERGENCY_FILING',
       'EMPLOYER_COUNTRY', 'TYPE_OF_REPRESENTATION',
       '790A_ADDENDUM_B_ATTACHED', 'WORK_CONTRACTS_ATTACHED',
       'EMPLOYER_MSPA_ATTACHED', 'SURETY_BOND_ATTACHED',
       'HOUSING_TRANSPORTATION', 'APPENDIX_A_ATTACHED',
       'JOINT_EMPLOYER_APPENDIX_A_ATTACHED', 'TOTAL_WORKERS_NEEDED',
       'TOTAL_WORKERS_H2A_REQUESTED', 'TOTAL_WORKERS_H2A_CERTIFIED',
       'EMPLOYMENT_BEGIN_DATE', 'EMPLOYMENT_END_DATE', 'ON_CALL_REQUIREMENT',
       'ANTICIPATED_NUMBER_OF_HOURS', 'WAGES_PER_HOUR',
       '790A_ADDENDUM_A_ATTACHED', 'EDUCATION_LEVEL', 'WORK_EXPERIENCE_MONTHS',
       'TRAINING_MONTHS', 'CERTIFICATION_REQUIREMENTS', 'DRIVER_REQUIREMENTS',
       'CRIMINAL_BACKGROUND_CHECK', 'DRUG_SCREEN',
       'ADDENDUM_B_WORKSITE_ATTACHED', 'ADDENDUM_B_HOUSING_ATTACHED',
       'ADDENDUM_C_ATTACHED'],
      dtype='object')


df_LC= pd.get_dummies(df['H2A_LABOR_CONTRACTOR'])
df= pd.concat([df_LC, df], axis=1)
df= df.rename(columns= {'N' : 'H2A_LABOR_CONTRACTOR_NO',
                           'Y' : 'H2A_LABOR_CONTRACTOR_YES',})
df.head(2)


df_NTN= pd.get_dummies(df['NATURE_OF_TEMPORARY_NEED'])
df= pd.concat([df_NTN, df], axis=1)
df= df.rename(columns= {'Other Temporary Need' : 'OTHER_TEMPORARY_NEED',
                           'Seasonal' : 'SEASONAL',})
df.head(2)


df_EF= pd.get_dummies(df['EMERGENCY_FILING'])
df= pd.concat([df_EF, df], axis=1)
df= df.rename(columns= {'N' : 'EMERGENCY_FILING_NO',
                           'Y' : 'EMERGENCY_FILING_YES',})
df.head(2)


df_ABA= pd.get_dummies(df['790A_ADDENDUM_B_ATTACHED'])
df= pd.concat([df_ABA, df], axis=1)
df= df.rename(columns= {'Y' : '790A_ADDENDUM_B_ATTACHED_YES',
                           'N' : '790A_ADDENDUM_B_ATTACHED_NO',})
df.head(2)


df_AAA= pd.get_dummies(df['790A_ADDENDUM_A_ATTACHED'])
df= pd.concat([df_AAA, df], axis=1)
df= df.rename(columns= {'Y' : '790A_ADDENDUM_A_ATTACHED_YES',
                           'N' : '790A_ADDENDUM_A_ATTACHED_NO',})
df.head(2)


df_WCA= pd.get_dummies(df['WORK_CONTRACTS_ATTACHED'])
df= pd.concat([df_WCA, df], axis=1)
df= df.rename(columns= {'Y' : 'WORK_CONTRACTS_ATTACHED_YES',
                           'N' : 'WORK_CONTRACTS_ATTACHED_NO',})
df.head(2)


df_SBA= pd.get_dummies(df['SURETY_BOND_ATTACHED'])
df= pd.concat([df_SBA, df], axis=1)
df= df.rename(columns= {'Y' : 'SURETY_BOND_ATTACHED_YES',
                           'N' : 'SURETY_BOND_ATTACHED_NO',})
df.head(2)


df_HT= pd.get_dummies(df['HOUSING_TRANSPORTATION'])
df= pd.concat([df_HT, df], axis=1)
df= df.rename(columns= {'Y' : 'HOUSING_TRANSPORTATION_YES',
                           'N' : 'HOUSING_TRANSPORTATION_NO',})
df.head(2)


df_AAA= pd.get_dummies(df['APPENDIX_A_ATTACHED'])
df= pd.concat([df_AAA, df], axis=1)
df= df.rename(columns= {'Y' : 'APPENDIX_A_ATTACHED_YES',
                           'N' : 'APPENDIX_A_ATTACHED_NO',})
df.head(2)


df_JE_AAA= pd.get_dummies(df['JOINT_EMPLOYER_APPENDIX_A_ATTACHED'])
df= pd.concat([df_JE_AAA, df], axis=1)
df= df.rename(columns= {'Y' : 'JOINT_EMPLOYER_APPENDIX_A_ATTACHED_YES',
                           'N' : 'JOINT_EMPLOYER_APPENDIX_A_ATTACHED_NO',})
df.head(2)


df_OCR= pd.get_dummies(df['ON_CALL_REQUIREMENT'])
df= pd.concat([df_OCR, df], axis=1)
df= df.rename(columns= {'Y' : 'ON_CALL_REQUIREMENT_YES',
                           'N' : 'ON_CALL_REQUIREMENT_NO',})
df.head(2)


df_CR= pd.get_dummies(df['CERTIFICATION_REQUIREMENTS'])
df= pd.concat([df_CR, df], axis=1)
df= df.rename(columns= {'Y' : 'CERTIFICATION_REQUIREMENTS_YES',
                           'N' : 'CERTIFICATION_REQUIREMENTS_NO',})
df.head(2)


df_DR= pd.get_dummies(df['DRIVER_REQUIREMENTS'])
df= pd.concat([df_DR, df], axis=1)
df= df.rename(columns= {'Y' : 'DRIVER_REQUIREMENTS_YES',
                           'N' : 'DRIVER_REQUIREMENTS_NO',})
df.head(2)


df_CBC= pd.get_dummies(df['CRIMINAL_BACKGROUND_CHECK'])
df= pd.concat([df_CBC, df], axis=1)
df= df.rename(columns= {'Y' : 'CRIMINAL_BACKGROUND_CHECK_YES',
                           'N' : 'CRIMINAL_BACKGROUND_CHECK_NO',})
df.head(2)


df_DS= pd.get_dummies(df['DRUG_SCREEN'])
df= pd.concat([df_DS, df], axis=1)
df= df.rename(columns= {'Y' : 'DRUG_SCREEN_YES',
                           'N' : 'DRUG_SCREEN_CHECK_NO',})
df.head(2)


df_ABWA= pd.get_dummies(df['ADDENDUM_B_WORKSITE_ATTACHED'])
df= pd.concat([df_ABWA, df], axis=1)
df= df.rename(columns= {'Y' : 'ADDENDUM_B_WORKSITE_ATTACHED_YES',
                           'N' : 'ADDENDUM_B_WORKSITE_ATTACHED_NO',})
df.head(2)


df_ABHA= pd.get_dummies(df['ADDENDUM_B_HOUSING_ATTACHED'])
df= pd.concat([df_ABHA, df], axis=1)
df= df.rename(columns= {'Y' : 'ADDENDUM_B_HOUSING_ATTACHED_YES',
                           'N' : 'ADDENDUM_B_HOUSING_ATTACHED_NO',})
df.head(2)


df_ACA= pd.get_dummies(df['ADDENDUM_C_ATTACHED'])
df= pd.concat([df_ACA, df], axis=1)
df= df.rename(columns= {'Y' : 'ADDENDUM_C_ATTACHED_YES',
                           'N' : 'ADDENDUM_C_ATTACHED_NO',})
df.head(2)


df_ACA= pd.get_dummies(df['TYPE_OF_EMPLOYER_APPLICATION'])
df= pd.concat([df_ACA, df], axis=1)
df.head(2)


mode_val = df['TYPE_OF_REPRESENTATION'].mode()[0]
df['TYPE_OF_REPRESENTATION'] = df['TYPE_OF_REPRESENTATION'].fillna(mode_val)


#replace values with 1s and 0s

df['TYPE_OF_REPRESENTATION'] = df['TYPE_OF_REPRESENTATION'].replace({'Agent':1,'Attorney':0})


df_ACA= pd.get_dummies(df['EMPLOYER_COUNTRY'])
df= pd.concat([df_ACA, df], axis=1)
df.head(2)


df.drop('OTHERS',axis=1,inplace=True)


del df['H2A_LABOR_CONTRACTOR']
del df['NATURE_OF_TEMPORARY_NEED']
del df['EMERGENCY_FILING']
del df['790A_ADDENDUM_A_ATTACHED']
del df['WORK_CONTRACTS_ATTACHED']
del df['HOUSING_TRANSPORTATION']
del df['APPENDIX_A_ATTACHED']
del df['JOINT_EMPLOYER_APPENDIX_A_ATTACHED']
del df['ON_CALL_REQUIREMENT']
del df['790A_ADDENDUM_B_ATTACHED']
del df['CERTIFICATION_REQUIREMENTS']
del df['DRIVER_REQUIREMENTS']
del df['CRIMINAL_BACKGROUND_CHECK']
del df['DRUG_SCREEN']
del df['ADDENDUM_B_WORKSITE_ATTACHED']
del df['ADDENDUM_B_HOUSING_ATTACHED']
del df['ADDENDUM_C_ATTACHED']
del df['TYPE_OF_EMPLOYER_APPLICATION']
del df['EMPLOYER_COUNTRY']
del df['TYPE_OF_REPRESENTATION']
del df['EMPLOYER_MSPA_ATTACHED']
del df['SURETY_BOND_ATTACHED']
del df['TOTAL_WORKERS_NEEDED']
del df['CASE_NUMBER']
del df['TOTAL_WORKERS_H2A_CERTIFIED']
del df['ANTICIPATED_NUMBER_OF_HOURS']


df.shape

(17753, 52)


df.head()


df['EMPLOYMENT_BEGIN_DATE'] = pd.to_datetime(df['EMPLOYMENT_BEGIN_DATE'], format="%d-%m-%Y")
df['EMPLOYMENT_END_DATE'] = pd.to_datetime(df['EMPLOYMENT_END_DATE'], format="%d-%m-%Y")


df['EMPLOYEMENT_DAYS'] = (df['EMPLOYMENT_END_DATE'] - df['EMPLOYMENT_BEGIN_DATE']).dt.days


del df['EMPLOYMENT_END_DATE']
del df['EMPLOYMENT_BEGIN_DATE']


df['WAGES_PER_HOUR'] = df['WAGES_PER_HOUR'].astype(float)


from imblearn.combine import SMOTETomek
from imblearn.under_sampling import TomekLinks
import plotly.express as px
import plotly.io as pio
import plotly.graph_objects as go
pio.templates.default = "plotly_white"


X = df.drop('CASE_STATUS', axis=1)  # Replace 'target_column' with the name of your target column
y = df['CASE_STATUS']


X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)


segment_counts = df['CASE_STATUS'].value_counts()
pastel_colors = px.colors.qualitative.Pastel
fig_before = go.Figure(data=[go.Bar(x=segment_counts.index, y=segment_counts.values,marker=dict(color=pastel_colors))])

fig_before


resample=SMOTETomek(tomek=TomekLinks(sampling_strategy='majority'))

from imblearn.over_sampling import SMOTE

strategy = {0:15538, 1:15538, 2:15538, 3:15538}
oversample = SMOTE(sampling_strategy=strategy)
X, y = oversample.fit_resample(X, y)
#X, y = resample.fit_resample(X, y)


segment_counts = y.value_counts()
fig_after = go.Figure(data=[go.Bar(x=segment_counts.index, y=segment_counts.values,marker=dict(color=pastel_colors))])

fig_after


model_dt = DecisionTreeClassifier(random_state=42)
model_dt.fit(X_train, y_train)

DecisionTreeClassifier(random_state=42)

DecisionTreeClassifier(random_state=42)


y_pred_dt = model_dt.predict(X_test)
accuracy = accuracy_score(y_test, y_pred_dt)
print(f"Accuracy: {accuracy:.3f}")

Accuracy: 0.920


from sklearn.ensemble import RandomForestClassifier


random_forest_model = RandomForestClassifier(n_estimators=100, random_state=42)
random_forest_model.fit(X_train, y_train)

RandomForestClassifier(random_state=42)

RandomForestClassifier(random_state=42)


y_pred_rf = random_forest_model.predict(X_test)
accuracy_rf = accuracy_score(y_test, y_pred_rf)
print(f"Random Forest Accuracy: {accuracy_rf:.3f}")

Random Forest Accuracy: 0.945


# Adjust class labels to start from 0
df['CASE_STATUS'] = df['CASE_STATUS'] - df['CASE_STATUS'].min()

# Split the dataset into features (X) and the target variable (y)
X = df.drop('CASE_STATUS', axis=1)
y = df['CASE_STATUS']


import xgboost as xgb

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)


model_xgb = xgb.XGBClassifier(objective='multi:softmax', num_class=4, random_state=42)


# Train the model on the training data
model_xgb.fit(X_train, y_train)

# Make predictions on the testing data
y_pred_xg = model_xgb.predict(X_test)


# Evaluate the model's accuracy
accuracy = accuracy_score(y_test, y_pred_xg)
print(f"Accuracy: {accuracy:.3f}")

Accuracy: 0.949


from sklearn.metrics import confusion_matrix
from sklearn.metrics import ConfusionMatrixDisplay
labels = ["Denied", "Withdrawn","Expired","Accepted"]


# Ensure the same number of samples
min_samples = min(len(X), len(y))
true_labels = X[:min_samples]
predicted_labels = y[:min_samples]


cm = confusion_matrix(y_test, y_pred_dt, labels=[0, 1, 2, 3])
disp = ConfusionMatrixDisplay(confusion_matrix=cm, display_labels=labels)
disp.plot()

<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x1ae7a8f4290>


from sklearn.metrics import classification_report

print(classification_report(y_test,y_pred_dt))

              precision    recall  f1-score   support

           0       0.11      0.17      0.13        23
           1       0.16      0.27      0.20        41
           2       0.77      0.77      0.77       368
           3       0.96      0.95      0.96      3119

    accuracy                           0.92      3551
   macro avg       0.50      0.54      0.52      3551
weighted avg       0.93      0.92      0.92      3551


cm = confusion_matrix(y_test, y_pred_rf, labels=[0, 1, 2, 3])
disp = ConfusionMatrixDisplay(confusion_matrix=cm, display_labels=labels)
disp.plot()

<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x1ae7ae3eb90>


from sklearn.metrics import classification_report

print(classification_report(y_test,y_pred_rf))

              precision    recall  f1-score   support

           0       0.00      0.00      0.00        23
           1       1.00      0.05      0.09        41
           2       0.89      0.72      0.79       368
           3       0.95      0.99      0.97      3119

    accuracy                           0.95      3551
   macro avg       0.71      0.44      0.46      3551
weighted avg       0.94      0.95      0.94      3551


cm = confusion_matrix(y_test, y_pred_xg, labels=[0, 1, 2, 3])
disp = ConfusionMatrixDisplay(confusion_matrix=cm, display_labels=labels)
disp.plot()

<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x1ae04f23b50>


from sklearn.metrics import classification_report

print(classification_report(y_test,y_pred_xg))

              precision    recall  f1-score   support

           0       0.17      0.04      0.07        23
           1       0.50      0.10      0.16        41
           2       0.86      0.80      0.83       368
           3       0.96      0.98      0.97      3119

    accuracy                           0.95      3551
   macro avg       0.62      0.48      0.51      3551
weighted avg       0.94      0.95      0.94      3551


from sklearn.model_selection import KFold #for K-fold cross validation
from sklearn.model_selection import cross_val_score #score evaluation
from sklearn.model_selection import cross_val_predict #prediction
kfold = KFold(n_splits=10, random_state=None) # k=10, split the data into 10 equal parts


# Decision Tree Classifier
score_dt = cross_val_score(DecisionTreeClassifier(), X, y, cv=10)
print(score_dt)
print("Avg :", np.average(score_dt))

[0.77815315 0.93468468 0.95551802 0.94535211 0.9543662  0.93577465
 0.91098592 0.85239437 0.82929577 0.77014085]
Avg : 0.8866665715010786


# Random Forest Score
score_rf=cross_val_score(RandomForestClassifier(n_estimators=100),X, y,cv=10)
print(score_rf)
print("Avg :",np.average(score_rf))

[0.85923423 0.96058559 0.97240991 0.95943662 0.96619718 0.96394366
 0.93859155 0.88169014 0.8771831  0.83830986]
Avg : 0.9217581842405785


# XGBoost
score_xg = cross_val_score(xgb.XGBClassifier(), X, y, cv=10)
print(score_xg)
print("Avg :", np.average(score_xg))

[0.82376126 0.96171171 0.97353604 0.96507042 0.96957746 0.97070423
 0.94422535 0.88788732 0.87830986 0.8315493 ]
Avg : 0.9206332952670981


from sklearn.model_selection import GridSearchCV


# Define the hyperparameters for each model
decision_tree_params = {'max_depth': [3, 5, 10]}

random_forest_params = {'n_estimators': [10, 20, 30],
                        'max_depth': [3, 5, 10],
                        'min_samples_split': [2, 5, 10]}

xgboost_params = {'n_estimators': [10, 20, 30],
                   'max_depth': [3, 5, 10],
                   'learning_rate': [0.01, 0.05, 0.1]}


# Create the models
decision_tree_clf = DecisionTreeClassifier()
random_forest_clf = RandomForestClassifier()
xgboost_clf = xgb.XGBClassifier()


# Create the grid search objects
decision_tree_gscv = GridSearchCV(decision_tree_clf, decision_tree_params, cv=10)
random_forest_gscv = GridSearchCV(random_forest_clf, random_forest_params, cv=10)
xgboost_gscv = GridSearchCV(xgboost_clf, xgboost_params, cv=10)


# Fit the models
decision_tree_gscv.fit(X, y)
random_forest_gscv.fit(X, y)
xgboost_gscv.fit(X, y)

GridSearchCV(cv=10,
             estimator=XGBClassifier(base_score=None, booster=None,
                                     callbacks=None, colsample_bylevel=None,
                                     colsample_bynode=None,
                                     colsample_bytree=None,
                                     early_stopping_rounds=None,
                                     enable_categorical=False, eval_metric=None,
                                     feature_types=None, gamma=None,
                                     gpu_id=None, grow_policy=None,
                                     importance_type=None,
                                     interaction_constraints=None,
                                     learning_rate=None...x_bin=None,
                                     max_cat_threshold=None,
                                     max_cat_to_onehot=None,
                                     max_delta_step=None, max_depth=None,
                                     max_leaves=None, min_child_weight=None,
                                     missing=nan, monotone_constraints=None,
                                     n_estimators=100, n_jobs=None,
                                     num_parallel_tree=None, predictor=None,
                                     random_state=None, ...),
             param_grid={'learning_rate': [0.01, 0.05, 0.1],
                         'max_depth': [3, 5, 10],
                         'n_estimators': [10, 20, 30]})

GridSearchCV(cv=10,
             estimator=XGBClassifier(base_score=None, booster=None,
                                     callbacks=None, colsample_bylevel=None,
                                     colsample_bynode=None,
                                     colsample_bytree=None,
                                     early_stopping_rounds=None,
                                     enable_categorical=False, eval_metric=None,
                                     feature_types=None, gamma=None,
                                     gpu_id=None, grow_policy=None,
                                     importance_type=None,
                                     interaction_constraints=None,
                                     learning_rate=None...x_bin=None,
                                     max_cat_threshold=None,
                                     max_cat_to_onehot=None,
                                     max_delta_step=None, max_depth=None,
                                     max_leaves=None, min_child_weight=None,
                                     missing=nan, monotone_constraints=None,
                                     n_estimators=100, n_jobs=None,
                                     num_parallel_tree=None, predictor=None,
                                     random_state=None, ...),
             param_grid={'learning_rate': [0.01, 0.05, 0.1],
                         'max_depth': [3, 5, 10],
                         'n_estimators': [10, 20, 30]})

XGBClassifier(base_score=None, booster=None, callbacks=None,
              colsample_bylevel=None, colsample_bynode=None,
              colsample_bytree=None, early_stopping_rounds=None,
              enable_categorical=False, eval_metric=None, feature_types=None,
              gamma=None, gpu_id=None, grow_policy=None, importance_type=None,
              interaction_constraints=None, learning_rate=None, max_bin=None,
              max_cat_threshold=None, max_cat_to_onehot=None,
              max_delta_step=None, max_depth=None, max_leaves=None,
              min_child_weight=None, missing=nan, monotone_constraints=None,
              n_estimators=100, n_jobs=None, num_parallel_tree=None,
              predictor=None, random_state=None, ...)

XGBClassifier(base_score=None, booster=None, callbacks=None,
              colsample_bylevel=None, colsample_bynode=None,
              colsample_bytree=None, early_stopping_rounds=None,
              enable_categorical=False, eval_metric=None, feature_types=None,
              gamma=None, gpu_id=None, grow_policy=None, importance_type=None,
              interaction_constraints=None, learning_rate=None, max_bin=None,
              max_cat_threshold=None, max_cat_to_onehot=None,
              max_delta_step=None, max_depth=None, max_leaves=None,
              min_child_weight=None, missing=nan, monotone_constraints=None,
              n_estimators=100, n_jobs=None, num_parallel_tree=None,
              predictor=None, random_state=None, ...)


model_params = {
    'decision_tree': {
        'model': DecisionTreeClassifier(),
        'params' : {'max_depth': [3, 5, 10]
        }  
    },
    'random_forest': {
        'model': RandomForestClassifier(),
        'params' : {
            'n_estimators': [10, 20, 30],
            'max_depth': [3, 5, 10],
            'min_samples_split': [2, 5, 10]
        }
    },
    'xgboost' : {
        'model': xgb.XGBClassifier(),
        'params': {
            'n_estimators': [10, 20, 30],
            'max_depth': [3, 5, 10],
            'learning_rate': [0.01, 0.05, 0.1]
        }
    }
}


scores = []

for model_name, mp in model_params.items():
    clf =  GridSearchCV(mp['model'], mp['params'], cv=5, return_train_score=False)
    clf.fit(X,y)
    scores.append({
        'model': model_name,
        'best_score': clf.best_score_,
        'best_params': clf.best_params_
    })
    
df = pd.DataFrame(scores,columns=['model','best_score','best_params'])
df


xgboost_best = xgboost_gscv.best_estimator_

imp_df = pd.DataFrame({
    "Varname": X_train.columns,
    "Imp": xgboost_best.feature_importances_
})
imp_df.sort_values(by="Imp", ascending=False)

	CASE_STATUS	DAYS	TOTAL_WORKERS_NEEDED	TOTAL_WORKERS_H2A_REQUESTED	TOTAL_WORKERS_H2A_CERTIFIED	ANTICIPATED_NUMBER_OF_HOURS	WAGES_PER_HOUR	EDUCATION_LEVEL	WORK_EXPERIENCE_MONTHS	TRAINING_MONTHS
count	17753.000000	17753.000000	17753.000000	17753.000000	17367.000000	17744.000000	17753.000000	17753.000000	17753.000000	17753.000000
mean	2.845153	27.488593	21.741058	17.538782	17.409397	44.215681	83.415359	1.021912	2.410916	0.033628
std	0.455756	13.912720	69.628631	53.158573	51.688077	7.417920	385.636911	0.151691	2.138582	0.863319
min	0.000000	0.000000	1.000000	1.000000	1.000000	30.000000	0.080000	1.000000	0.000000	0.000000
25%	3.000000	18.000000	2.000000	2.000000	2.000000	40.000000	13.880000	1.000000	1.000000	0.000000
50%	3.000000	27.000000	5.000000	5.000000	5.000000	40.000000	15.660000	1.000000	3.000000	0.000000
75%	3.000000	35.000000	14.000000	12.000000	12.000000	48.000000	17.330000	1.000000	3.000000	0.000000
max	3.000000	790.000000	2100.000000	2100.000000	2100.000000	105.000000	7000.000000	5.000000	72.000000	100.000000

Importing all the essential libraries¶

Reading the dataset¶

Creating dummy variables for all the necessary categorical variables¶

Deleting all the original columns since they have been assigned dummy values¶

Checking empty values¶

Oversampling¶

Predictive Models¶

Decision Tree¶

Random Forest¶

XGBoost¶

Evaluation of the Model¶

Confusion Matrix¶

K-Fold Evaluation¶

Hyperparameter Tuning¶

	CASE_NUMBER	CASE_STATUS	DAYS	TYPE_OF_EMPLOYER_APPLICATION	H2A_LABOR_CONTRACTOR	NATURE_OF_TEMPORARY_NEED	EMERGENCY_FILING	EMPLOYER_COUNTRY	TYPE_OF_REPRESENTATION	790A_ADDENDUM_B_ATTACHED	...	EDUCATION_LEVEL	WORK_EXPERIENCE_MONTHS	CERTIFICATION_REQUIREMENTS	DRIVER_REQUIREMENTS	CRIMINAL_BACKGROUND_CHECK	DRUG_SCREEN	ADDENDUM_B_WORKSITE_ATTACHED	ADDENDUM_B_HOUSING_ATTACHED	ADDENDUM_C_ATTACHED
0	H-300-22231-426336	3	38	Individual Employer	N	Seasonal	N	UNITED STATES OF AMERICA	Agent	N	...	1	0	N	N	N	N	N	N	Y
1	H-300-22223-410500	2	48	Individual Employer	N	Seasonal	N	UNITED STATES OF AMERICA	Agent	N	...	1	3	N	Y	N	N	N	Y	N
2	H-300-22223-410821	2	52	Individual Employer	Y	Seasonal	N	UNITED STATES OF AMERICA	Attorney	Y	...	1	3	Y	Y	Y	Y	Y	Y	Y
3	H-300-22230-423268	2	20	Individual Employer	N	Seasonal	N	UNITED STATES OF AMERICA	Agent	N	...	1	3	N	Y	N	N	N	N	N
4	H-300-22230-423991	2	42	Joint Employer	N	Seasonal	N	UNITED STATES OF AMERICA	Agent	N	...	1	3	N	N	N	N	Y	Y	Y

	UNITED STATES OF AMERICA	Individual Employer	Joint Employer	ADDENDUM_C_ATTACHED_NO	ADDENDUM_C_ATTACHED_YES	ADDENDUM_B_HOUSING_ATTACHED_NO	...	H2A_LABOR_CONTRACTOR_YES	CASE_STATUS	DAYS	TOTAL_WORKERS_H2A_REQUESTED	EMPLOYMENT_BEGIN_DATE	EMPLOYMENT_END_DATE	WAGES_PER_HOUR	EDUCATION_LEVEL	WORK_EXPERIENCE_MONTHS
0	1	1	0	0	1	1	...	0	3	38	6	01-11-2022	28-08-2023	12.45	1	0
1	1	1	0	1	0	0	...	0	2	48	6	25-10-2022	01-06-2023	16.47	1	3
2	1	1	0	0	1	0	...	1	2	52	21	19-10-2022	10-04-2023	12.41	1	3
3	1	1	0	1	0	1	...	0	2	20	1	01-11-2022	31-03-2023	12.45	1	3
4	1	0	1	0	1	0	...	0	2	42	30	01-11-2022	30-06-2023	15.66	1	3

	model	best_score	best_params
0	decision_tree	0.877713	{'max_depth': 3}
1	random_forest	0.913870	{'max_depth': 5, 'min_samples_split': 10, 'n_e...
2	xgboost	0.922263	{'learning_rate': 0.1, 'max_depth': 3, 'n_esti...

	Varname	Imp
23	JOINT_EMPLOYER_APPENDIX_A_ATTACHED_NO	0.493285
49	EMPLOYEMENT_DAYS	0.260474
45	WAGES_PER_HOUR	0.075666
41	H2A_LABOR_CONTRACTOR_NO	0.035676
43	DAYS	0.023690
39	OTHER_TEMPORARY_NEED	0.011281
27	HOUSING_TRANSPORTATION_NO	0.011104
17	DRIVER_REQUIREMENTS_NO	0.010731
37	EMERGENCY_FILING_NO	0.010365
47	WORK_EXPERIENCE_MONTHS	0.009851
7	ADDENDUM_C_ATTACHED_NO	0.009289
29	SURETY_BOND_ATTACHED_NO	0.009132
33	790A_ADDENDUM_A_ATTACHED_NO	0.008672
48	TRAINING_MONTHS	0.008594
15	CRIMINAL_BACKGROUND_CHECK_NO	0.007811
9	ADDENDUM_B_HOUSING_ATTACHED_NO	0.007110
13	DRUG_SCREEN_CHECK_NO	0.005828
44	TOTAL_WORKERS_H2A_REQUESTED	0.001443
38	EMERGENCY_FILING_YES	0.000000
36	790A_ADDENDUM_B_ATTACHED_YES	0.000000
35	790A_ADDENDUM_B_ATTACHED_NO	0.000000
34	790A_ADDENDUM_A_ATTACHED_YES	0.000000
40	SEASONAL	0.000000
42	H2A_LABOR_CONTRACTOR_YES	0.000000
32	WORK_CONTRACTS_ATTACHED_YES	0.000000
31	WORK_CONTRACTS_ATTACHED_NO	0.000000
30	SURETY_BOND_ATTACHED_YES	0.000000
46	EDUCATION_LEVEL	0.000000
0	CANADA	0.000000
28	HOUSING_TRANSPORTATION_YES	0.000000
12	ADDENDUM_B_WORKSITE_ATTACHED_YES	0.000000
2	Association - Agent	0.000000
3	Association - Joint Employer	0.000000
4	Association - Sole Employer	0.000000
5	Individual Employer	0.000000
6	Joint Employer	0.000000
8	ADDENDUM_C_ATTACHED_YES	0.000000
10	ADDENDUM_B_HOUSING_ATTACHED_YES	0.000000
11	ADDENDUM_B_WORKSITE_ATTACHED_NO	0.000000
14	DRUG_SCREEN_YES	0.000000
26	APPENDIX_A_ATTACHED_YES	0.000000
16	CRIMINAL_BACKGROUND_CHECK_YES	0.000000
18	DRIVER_REQUIREMENTS_YES	0.000000
19	CERTIFICATION_REQUIREMENTS_NO	0.000000
20	CERTIFICATION_REQUIREMENTS_YES	0.000000
21	ON_CALL_REQUIREMENT_NO	0.000000
22	ON_CALL_REQUIREMENT_YES	0.000000
24	JOINT_EMPLOYER_APPENDIX_A_ATTACHED_YES	0.000000
1	UNITED STATES OF AMERICA	0.000000
25	APPENDIX_A_ATTACHED_NO	0.000000